8

The anti-taser blog, Truth Not Tasers lists over 768 people that they claim have been killed by police tasers since the 1983.

On the other hand, the Wikipedia article contains an unsourced statement:

Police departments counter that while Tasers were used to subdue these individuals, their in-custody deaths were unrelated to their encounter, and could have likely been caused by more traditional police impact weapons (like batons).

Further, Taser International CEO Rick Smith claims is reported to have claimed:

police surveys show the device has saved 75,000 lives.

Does the use of tasers by the police lead to an increase in police killings (as implied by the linked blog), or does it lead to a decrease (as claimed by the Taser International CEO)?

In response to @JoeWreschnig's comment, the saved lives comment might refer to shooting with a taser instead of shooting with a gun. The relevant subquestions are therefore:

  • What is the conditional probability of being killed when being shot by a taser?
  • What is the conditional probability of being killed when being shot by a police gun?
  • In what fraction of lethal taser shots was the taser shot used instead of a gun shot?
  • In what fraction of lethal taser shots were other alternatives available?

I think that with these sub-questions, the main question can be answered.

Improve this question
10
  • 1
    "Have police tasers killed many hundreds of people in the USA? Or do they rather save lives?" These statements are not mutually exclusive. (Although I think it's questionable to consider "don't kill as often as guns" as "saving lives" - by this definition, nearly everything in the world is constantly saving lives.)
    – user792
    Commented Dec 17, 2012 at 12:27
  • the number killed is over a period of 20-30 years, if you average that out its about the same or less than lightning strikes.
    – Ryathal
    Commented Dec 17, 2012 at 13:29
  • 1
    @gerrit I am saying your question reads like its trying to create a problem where none actually exists by portraying something that kills at a rate no greater than random chance as an evil death machine. it also answers your first question, as a no tasers don't kill many hundreds, unless you pick a unit of time much larger than the standard year.
    – Ryathal
    Commented Dec 17, 2012 at 14:07
  • 2
    BTW. they are no longer referred to as "non-lethal", but as "less-lethal".
    – vartec
    Commented Dec 17, 2012 at 14:32
  • 1
    Note that taser may be substituted for batons rather than for firearms. Or for talking the subject down; or for self-discipline on the part of an officer who's pride has been tweaked. In any case it is not clear that the comparison is opening fire vs. taser. In principle taser should give officers a better selection of low-lethality responses to tough situations, but who knows how well that potential is realized on the ground. Certainly the worst cases of taser misuse reported in the news are fairly horrifing.
    – dmckee --- ex-moderator kitten
    Commented Dec 17, 2012 at 14:55

1 Answer 1

Reset to default
11

The CNN report refers to a 2009 report from the Police Executive Research Forum as a reference.

Chasing that down, it appears the PERF report is a September 2009 report to the National Institute of Justice, that goes by the clumsy name:

In this title "Conducted Energy Devices" is the generic name for Tasers. It was "quasi-experimental" in that they used matched controls from existing police databases, comparing seven police departments ("Law Enforcement Agencies", LEAs) with tasers to six without, and also for the two years before the tasers were introduced.

They controlled for:

  • force used by officer,
  • time frame of incident,
  • suspect race/gender/age,
  • suspect resistant behavior,
  • suspect weapon use,
  • agency policy on CEDs
  • size/density of LEA
  • population density for jurisdiction.

They found that CEDs improved the following measures:

  • Officer injuries
  • Officers receiving injuries requiring medical attention,
  • Suspect injuries
  • Suspect severe injuries
  • Suspects receiving injuries requiring medical attention, and
  • Suspects receiving an injury that resulted in their being sent to a hospital or other medical facility

They found no difference with:

  • number of suspect deaths
  • officer severe injuries,
  • officer injuries requiring hospitalization

Note: Crime statistics notoriously suffer from reporting biases - those doing the reporting are being judged on their data - which may make it more difficult to completely trust any study in the field.

Improve this answer
4
  • Are the Police Executive Research Forum a sufficiently independent source for this purpose?
    – gerrit
    Commented Dec 18, 2012 at 15:32
  • How about reading the study that you cite? The amount of subjects during the time period they used CED was double than in the timeperiod where they had no CEDs while the non-CED sites had the same amount of deaths in both periods. They make a few arguments of why that doubling of deaths is just random noise even through p<.05, but taking this study as evidence against increase in number of suspect deaths seems very fishy to me.
    – Christian
    Commented Dec 19, 2012 at 19:14
  • @Christian: Please be nice! You're right: they found a slightly significant difference in the raw data, but given the no of comparisons they made, that's not enough to draw a conclusion. They argued there's a confounding factor (the CED sites had fewer suspect deaths before deployment) and did an (admittedly low-powered) analysis to support that. Do you agree that, even if you believe that difference to be significant, both claims ("CEDs cause MORE deaths" versus "studies (presumably this one) show CEDs have saved 75,000 lives") are not supported. Do you think this warrants being included?
    – Oddthinking ♦
    Commented Dec 19, 2012 at 21:17
  • @Oddthinking: It depends what you mean by "supported". I don't think this evidence should make you think that either of the claim true. For me the claim "CEDs cause MORE deaths" is more likely to be true than before I read the data. | There also the issue that they analysed the number of suspect deaths with bivariate modelling instead of using multivariable modelling. This means that it's not controlled against the factors that you list. Why they didn't do that is a mystery to me.
    – Christian
    Commented Dec 20, 2012 at 15:27

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .